The geometry of the somata and dendritic trees of motoneurons innervating neck and shoulder muscles was investigated by using intracellular injections of HRP. In general, these motoneurons did not belong to a homogeneous population of motoneurons. Differences in average primary dendritic diameter, number of primary dendrites, and other measures of dendritic tree size were found between different neck and shoulder motoneuron groups. Several indices of proximal dendritic tree size (number of primary dendrites, sum of dendritic diameters, Rall's dendritic trunk parameter, and the sum of dendritic holes) were weakly correlated with the diameter or surface area of the soma. Some of these correlations depended on the muscle supplied by the motoneuron. The total combined dendritic length ranged from 66,660 to 95,390 microns. There was a weak, but positive, correlation between the diameter of primary dendrites and combined dendritic length. This relationship varied from motoneuron to motoneuron. The diameters of all dendrites of three trapezius motoneurons were examined in detail. The total dendritic surface area examined ranged from 415,000 to 488,100 microns 2 and represented approximately 99% of the total neuronal surface area. Last-order dendrites showed a high degree (39.9%) of taper. Dendritic tapering, by itself, was a major factor in the decrease of the (sum of dendritic diameters)3/2 measured at progressively distal sites from the soma. Although few parent and daughter dendrites obeyed the "three-halves law," the average exponent was 1.57. The diameters of primary dendrites and dendritic surface area were weakly correlated. The correlation between dendritic diameter and combined dendritic length or surface area improved if the weighted average of the diameter of second-order dendrites was used as a measure of dendrite size. Second-order dendrites, whose branches terminated in different regions of the spinal cord, showed different relationships between dendritic diameter and combined dendritic length or surface area. Comparisons between the motoneurons examined in the present study and motoneurons innervating other muscles indicate that, although all spinal motoneurons share several common features (e.g., long dendrites, dendritic tapering), each motoneuron group has a set of unique features (e.g., soma shape, relationship between primary dendrite diameter and dendritic surface area). Thus, the rules governing motoneuron dendritic geometry are not fixed but depend on the species of the motoneuron.